John Motley | How AI Can Help Shippers Find Their Happy Path

[MUSIC] Hello and welcome to the Freightvine podcast, your source for all things freight transportation. I'm Chris Campos, Chief Scientist at DAT Freight Analytics. And today my guest is John Motley, the Founder and CEO of Lognet. Now Lognet was founded over 30 years ago. In our conversation, John and I discuss the evolution of supply chain technology over that time period. John recounts Lognet's origins as an early pioneer, being the first client service solution to utilize SQL databases on PCs. And they focus specifically on visibility across order, transportation, and warehouse management. Today, Lognet's core technology centers on large trade models, which they are very similar to large language models, but based on graph theory with attributes like weight, cube, and other carrier data. We also discuss how AI is replacing traditional user interfaces by fostering a conversational and agentic solution and moving past the aesthetically pleasing but quite often useless features like maps. This episode is essential for anyone wanting to understand how importers and retailers can leverage advanced technology to mitigate volatility from tariffs, geopolitical challenges, and other disruptions, and keep the majority of their freight on what John refers to as the happy path. Following my conversation with John, I'll present the latest truckload market update. So let's get started. Hi John, welcome to the freight find podcast. Thank you very much for inviting us. Yeah, I'll be honest, I've always heard of Lognet and I haven't interacted with it directly. And so I'm really interested finding out more and then get into some of the AI stuff that you guys are doing really interesting things with. But let's start. How did Lognet come into being? Because you've found it and have been with it for how many years now? 30 years now? Then a little over 30 years, yeah. Right. So how did it start? Why did it start? Well, I actually started as an engineer from the US Merchant Marine Academy and went into ships and then worked shoreside and thought after doing terminal operations and equipment control and various things, I could actually do something on the technical side. I didn't think you could make much money doing computers, but then polished off my MBA at NYU, had the privilege of having Dr. Deming there in his last couple years of teaching operations research. And ironically, you know, just after finishing our final, you learned some of the most advanced forecasting techniques in the world that you will never use. He thought that, you know, if your audience didn't understand the techniques, you weren't going to be able to use them. And it was more important to have metrics that people understood how they were derived. Oh my God. So that is the argument. I have this argument every day with people using advanced machine learning or AI black boxes where the MAP, the mean absolute percent error might go up by a point, but you can't explain anything. And it is the myopic forces of not wanting you to adopt something really have always impeded us for using more advanced techniques. Only recently we get, we see a ton of particularly large retailers in large importers, you're across the table from industrial engineers out of top schools, and they're ready to go with all kinds of advanced techniques. But really the machine learning is, I think, really taking off. And he would have been proud to take a look at the fact that a machine would be interpreting that and would have no problems doing integral analysis for non-linear solutions. Yeah. So, so this was 90, 192. And so the state of computers were still a little, that was like the IBM. We were the first platform to use SQL databases on PCs and in a client server, we were the first client server solution out there. We've always been focused on end-to-end sub-sq operations of visibility. So that's always been our reason for being. And across the order management, transportation management, warehouse management scope. So that end-to-end scope and with a hint at the, you know, international being our real swim lane. So, so let me make sure I understand end-to-end, that makes some sense. But what do you mean by sub-sq? That means that we know what's in the carton, we know the label, the whatever is identifying that carton, we know if it's a parallel, we know the size is the colors, the widths, the mentions, if it's electronics, this, what's the serial number. So, and we have continuity of that from end to end. So if we're going, and ironically, as the world is evolved from going, you know, factory floor to the customer door, that's really just been another evolution of our core model. So the idea is that you're not just the container number, you're knowing what's in it. So you're okay. It's not just tracking in one mode trying to piece together one or two modes, it's threading end to end. The fact that it started on a production line, went through some intermediate facility and then got transformed. It's going to have to go through multiple customs jurisdictions. All of that being part of the problem. So how did your original vision in 1992, your freshly graduated minted from Stern grad school, and you have this company, what was your vision then and how does that map to what it is now? It's essentially the same thing. As a matter of fact, we couldn't get patents or copyrights on a lot of the material because it wasn't, you know, there was all kinds of contention with method patents back in the day. So if you were to look at our trademark for Lognet, it reads like a patent. It's a gigantic description of network based, you know, because the internet was at its, it was pre-internet. So it's written as a network solution for trade management and all things and and documentation because we put into the problem though, may not just, so our model became eventually, I've always been a big follower of AI and I hate to overuse the term these days, but we've really built like a large trade model instead of a large language model, you know, natural language, sort of the be-all-end all of the neural networks behind LLMs, but we looked at what was the, what were we moving and what was about that that we would, we could make graphs of and then use a lot of graph theory which ended up being the same sorts of things that went into the LLMs about the products and things we were moving and what would that look like. And so that's the original concept and it's grown as very much analogous to a lot of the different disciplines of AI have evolved. We've been strong components of each one of them. So it sounds like the vision has been there to try to, you know, at the finest grain level end to end, but technology has had a sea change since then because you use the client server word and I can't think of the last time I heard client server, I remember, yeah, after PNG and then had to be on Windows NT. Yes, we survived all of that, we transitioned from that client server into internet based into internet through the dot com bubble over the hill and round to we have a slightly different vision for internet hosting and how to do that right. Most of the KPIs we have because we're a execution platform as the latency and the downtime standards are extremely high. So we couldn't afford an outage like on AWS because do you host your own? Yes, we have data centers that have our own hardware. Lognet hosts those not that do you have any customers that have their own servers in their basement somewhere or is that those days? No, those days are gone. Do you find them curious? Because I got out of software and that was just coming in and remember of the phrase application service providers, ASPs, that was the first thing. Yes, they were looked down upon because they weren't that, you know, lipstick on a pig was the phrase that I heard a lot trying to sell that, but it's changed a lot. The requirements are extraordinary. The security, the levels of security, the amount of partners you have to have in security, uptime, how you manage that uptime, all the different regulations around GPR, the European Union that you have to consider, sock compliance. All those things are you're really relying on someone else to take care of and the exposure, but the breaking point for us was most of the commitments on uptime just aren't good enough. You know, if our systems down, trucks aren't moving, ships aren't, things aren't making it on to ships, stuff isn't coming out of factories, they can't stand that kind of an outage. So, help talk me through what the solution looks like today, just at a high level, how does it span and how has that changed over the last several years? So, we have, we handle the moment that you negotiate a purchase order, typically an international order, place it with factory, we dialogue with the factory, do work and process management, take it through quality, into packing, we generate labels for them to place on the outer cartons and the inner cartons, they scan them into a conveyance that goes on a plane or boat someplace, we manage that flow out of customs of the exporting country, into the importing country, all the documentation that goes with that, preparing delivery orders, we can go into translating and transshipment facilities where we might segregate something out. A very common process is for us to get manufacturing orders where they're aggregate orders that are then moved into either regional slots or ports of entry that we then get blasted in, sales orders that distribute those. So people swing the doors open and it left as a Los Angeles container and they opened it up and it's got 50 different destinations in it now because we've allocated all the product. So we're like a warehouse that's virtual around the world. - So that the final delivery destination can change in route. - Yes. So let me make sure I understand. So that's some of a manufacturer of whatever widgets and a retailer and so the parts are moving from my plant, name it wherever is in the Pacific Rim, coming into the US, who owns Lognet, the manufacturer or the retailer managing it or the third party transportation providers? - We're agnostic. We will be an equal opportunity offender and work with all parties involved, whoever would like to, more and more is becoming the retailer. We're seeing them go up the supply chain and really take over quite a bit. It helps them from a first-cost basis. So if I'm entering the country and I pay the duty on goods that I purchased, F.O.B. at the port of Yantan, China, that means from the production facility, somewhere in southern China, all the way to the port, all the costs that accrue are being embedded into the product by taking that F.O.B. point, owning that as the retailer or importer and then backing up and saying, all right, well, I'm gonna pay for the consolidation at the port. I'm gonna pay for trucking to the port and backing that all the way up to the factory door, all the costs that I take out of that, I can reduce my commercial invoice into customs and I can abate my customs duties by going all the way back to what's known as first-cost. - That makes a ton of sense. It seems like has that always been or when you first started to go the other way? - Absolutely, it still is today a wall. So the third party logistics providers, the factories, they're all basically saying, don't worry about taxes or complexity of dealing in a foreign currency. I have got my boots on the ground, we'll take care of everything, we'll just get it on a ship for you and it all will be good. That's why your terms of sale should be F.O.B. Not X works. The move, two X works, they say, oh, it's so complicated that with the AI platforms, that's what's really the meat of what's changing today, especially with the tariff burden that people are seeing. It gives them a path to really mitigate some of that cost. - Yeah, 'cause some of the complexity now can be handled with some of the newer technology and we'll get to that in a second but I wanted to ask some about tariffs 'cause you mentioned the T word first. Wait, you obviously 2025 has been very volatile on the statistic I like to give is, as of the end of July, a new tariff announcement came out on average every nine days, every nine days, whether really added a new one, repeal, change, whatever. How did that manifest itself in your lognet system? Did that increase usage? Did that, how did that stress your system? - So the top importers and retailers have already sort of taken their approach of wanting to have a de-risk supply chain. So they wanna go and say, no more than X percent of my sourcing is gonna be on country because there's too much risk. The fact that whether it's gonna be a geopolitical incident or a tariff incident or typhoon or hurricane, something is gonna happen and for that reason, there's gonna have to be some diversity. So some clients have really done a fantastic job in mitigating that tariff and/or geopolitical and/or whatever happens to be who would foresee the hoodies popping up and starting the bomb ships as they try to pass by the canal with a crazy world. - That's funny 'cause we just had an event up here at MIT on disruptions. How do you navigate disruptions? And that was one of the questions we asked, what strategy do you employ? And dual sourcing, dispersing, China plus one is the common one, but people were talking more about China plus N. And then you have it multiplied. - Everyone is way down that road already. That's definitely already happening tremendously and also having agility and facility. So one of the big things we've always promoted is the ability to connect to different systems because as you go and it's a supply chain. So who are you gonna hook into for the next link in your chain, you really don't know who that is, but having that facility is super important. And some platforms look at having a giant thing that everyone connects into, but really AI and agentech platforms in particular, the ability to integrate and collaborate has been far more important. - So that's really an interesting point and we'll talk more about AI and agentech, but 'cause there are two extremes. One is you have a centralized system, everyone has to play by those rules and plug in. The other is you have some separate systems but they talk to each other. They're able to communicate. And so you have the right rules of the road, I guess. Is that a fair way of saying it? - Yeah, and it's sort of, one thing we try to do is support the continuum of what's out there. You know, we take the position of it is where it is. And where it is may not be with the most advanced platform does gonna be a broad continuum of what the capabilities are there. It's gonna be, are they gonna have EDI, API? What are they gonna be able to communicate with? And even the digital folders, where when we have to integrate with them, they'll have us talk to people who have API, just a death and said, well, your web hook has to work and if it only did a good job. And we say, well, we set up a web hook, we read your payload and it's empty. So you're sending over four events. I need 32. This customer wants a little more fidelity than that. Can we get past the API speak and get into the business speak about where these events are gonna come from? - Yeah, that's, it's interesting. We've had discussion with different companies about EDI. And the beautiful thing about EDI is that everyone has their own standard, right? 'Cause they're all different. And so the question is, 'cause we struggle with this too at DAT, you get emails and you get so much unstructured data and things, how do you handle and determine the one truth? 'Cause a lot of times there's lags in EDI, APIs might come in differently. You might get a Geogarve, a ping for something. How do you handle that and determine what's the one, what's the real truth? - Well, schema management is huge and having a sort of a flexible core schema is something that we've always had as a must have, that to assume that you're gonna get an SAP interface that's gonna be the same as the last SAP is a bad position. The APIs are also horrific in what they provide. We try as much as possible to go with like, hey, we're using the American Trucking Association's 214. Glad you are too. You forgot three things that you're signatory for on that 214. And the DCSA on Ocean Shipping APIs is terrific to say, let's try to go and work with a standard instead of your proprietary interface because we gotta reuse this and get the cost down for integration. So we rely on standard organizations, we support standards organization, they help, but in no way do they, I wish there was a forum out there where people could maybe read it or something and say what really is happening? 20% of the trucking moves, for example, in a port are not automated. There's no telematics, it's somebody with a pencil. The data's coming in a day later. If it's a rail facility out in the Midwest, some driver is turning a log in up to 24, 48 hours later and it's not real time. So we have customers that are trying to ratchet down contractually how long after an operating event, they'll accept the EDI and trying to go under 24 hours, there is a lot of wailing and gnashing of teeth and we meet every single day with dozens of different providers and it gets really chippy when we start asking for events to hit the 24 hour period. - Chippy's the right word. - Yes. - You'll chippy. Yes, I can see that over the last 30 years, do you find the time to bring a new vendor on a new Transfacian Provider system? Has that decreased or has that increased? - It's decreased. Most can turn things on really quickly. There are as soon as you get into the purchase order, it's a mess. Things slow down dramatically. That's where schema goes and falls apart because every single purchase order is different and some of the, I think consolidators, people that have worked with purchase orders a long time, the larger companies, they're pretty good at standing things up but if they haven't, it's a chore. - Yeah, yeah. Let me ask another thing that's happened over the course of the year, the repeal of the Diminimus Exception. What kind of effect did that have across your customer base? Where the customs change from 800 to where you have to pay customs? - So that was really a mortal thing. If you had a lot of air freight with the team and the machines that are out there, it was really looking at their method of operating and could they compete with the mass retailers who were working with more of the traditional approach? So it really didn't impact as much. It's the same process. It's just either, it's more like a parcel management problem as opposed to a full container load problem and we deal with both. What it did force a change in was the operating change and a lot of the mass retailers thought, okay, this is my new problem and it gets back to that going X works because this is, this sheen was an entity that controlled the manufacturing and they did a brilliant job of saying, I'm gonna take these manufacturers. collaborate with them, get them all to a centralized airport and then consolidate there and ship out to get a decent cost on air freight. So that was a terrific model for X-Works shipping. And so I think a lot of our customers took a look at how could they adapt that and would adopt sort of the best of that, but blend it with their huge distribution networks so they could optimize the cost at both ends. >> Yeah, that makes sense. So let's not get to the reason why I reached out to you in the first place, 'cause I saw a quote from you about the talking about that AI is replacing UIs of user interfaces. And so can you explain what you meant by that? >> Yeah, so it was back in 2022. We were with the Journal of Commerce and we were talking about, you know, every, we had a competitor after a competitor talking about UX. And all we had, it was the bloody maps. Everyone's got a map in their interface and it looks beautiful and it's maybe got some blinking lights on it and it's all UX. >> But we know, we know map cell software. >> Yes. >> And it's in his maps. >> Yes, and that was it. >> You had, if you didn't have a map, forget it. And so we have a map, it's beautiful. It has all the continents on it. We can put a ship and watch it go across. But yesterday, today, tomorrow, it's only going maybe 30 miles an hour at best. It's gonna still be on the water. And so my point at the time was all this talk of UX, it was really a marketing talk as opposed to a substantive talk on, is it improving your estimated rival time and customer fulfillment? And what is the best form of customer experience? It's getting to a wider-- >> Do you think that maps, seeing those, why do you think that executives like those does it make them feel comfortable? It's like, I know UPS will show you where your package is and everything. Why do you think that it's such a, 'cause it's still out there. >> It's beautiful. >> I think they project scale. It's a little different when you put a map up and there's a ship out there. And maybe it's more aspirational. But when you put it up there and it's full of ships and planes and boats and trucks, it looks interesting and cool. And if you can coordinate that here, there's a cluster of problems, where there's something just blew up, the question being if you didn't know it blew up and had a look at a map to determine that. >> That's the bigger question. >> And so ironically, for me, the big thing was, why would you have a map showing that Los Angeles is backed up by 200 ships and your shipment is going to Los Angeles? If you had AI, you would have gone to Houston or Prince Rupert in Canada, railed it into the Midwest and only been two days late and gotten your product on the shelves. And so all of our customers did that. The AI routed into Round LA. Then they were more than happy to see that other customers were waiting online to get off the ships during COVID, but couldn't understand why would you do that? And to us, it was more proof of they're not using AI. >> It isn't. >> And that's the best. >> When you say AI here, for a lot of this, it sounds like that's almost the optimization routing. Is that what you mean by that so? When you say AI, what do you mean? >> So I mean, so probably more of the traditional touring model where you have machine learnings, a classification of AI. >> Got it. >> So some of your stochastic, those are more, as you well know, operations research and more procedural forms of AI. Where large language models to me would be more the class of a natural language interface as being the reason for their being and also how well they do reasoning and a solutioning and that they're getting better, but not really appropriate for doing the reasoning. You might want to rely more on a machine learning algorithm that was more procedural. And then the robotic process automation, robotics being a form of AI, but I sort of segregate them more along the traditional touring classifications of AI. >> And so a touring classification, as I understand it is where I'm having a conversation with an entity and I can't tell whether it's a computer or a person. That was one of their things and they had the, it had to have memory, it had to be able to communicate. And I think that the total touring test had, they added the ability to perceive things and to manipulate things. >> Does anyone arguing that the touring test has not been passed? >> I think we're way past it, yeah. >> Yeah, I think we're way through five. >> Especially when they, when you look at, so I mean, when you look at the accuracy of forecasting and I think as you look at how quick and the, used to have 24 hours to process the optimization of like an I-log on a big IBM mainframe. And now you have millisecond response where you're doing routing optimization that would take, you'd have to run everything in batch overnight. And now you can, and now you can combine those with the LLMs, you can augment the responses and get incredibly rich responses in real time. >> I think that's the real, that's what I see the most promising path is not like, let's forget all the OR stuff we did and let's go all, language, language, it's where they fit together. It's really interesting how they can come up. >> And, let me, >> That's the reason for an agentic solution. >> So, so when I heard AI replacing UIs, what I had in my mind is getting rid of a graphical interface and going making more conversational. Like, members start track, I remember one of the start track moves, or Scotty is there back, I'm and he's talking to the computer, and it doesn't respond. Is that kind of what you remember by the natural angle processing that will be more conversational interfaces? >> Absolutely, I think if you took, and that's an important, I think it's to understand our solution, if you took a first principles approach and you said, how would I build supply chain solution, you would throw out 99% of the solutions that are out there. In particular, in freight forwarding, it's built around 2000 person service centers and factories of people that know how to do, they can do an air freight quote, or a motor quote, or a spot quote, or a contract quote, and that person can do that in two seconds flat. And you put 400 people in a row that do those things and you can process it, and they're in, but it's a very cost effective way of doing business. We call it factory forwarding or factory brokerage. By doing that, it's extremely efficient. So whatever solution comes along to replace it, must be really efficient because you're starting out with a tough competitor. And that model, I think for us, if you take the first principles look, well, it should be able to do all these things in parallel. And you should be able to look at really a graph, a trade graph, not a language graph, that every single node knows about the more attributes the better. So when you think of large language models taking thousands and thousands of documents and putting all of those attributes and then tokenizing them to assemble a statistical method of coming back with, what's the sentence look like? Then that's what we did. If you took the weight, the cube, the day of the week, the carrier, the exponentially whatever you want to take in about that thing that you're moving and how it's moving, that graph is what Lognet is and it's hard. - So, Joe, when you say trade graph, 'cause I mean, people as soon as might not be experts at large language models, I think people are starting to understand pattern matching and things like that. What you're saying is your base token, is it a shipment? Is it a move? - It is, and it's everything about that move. So it means that not just the physical part, the document part, everything about it. The model, for example, that we had a customer that was having problems moving shipment through the port of Tacoma. And it was a shipments of tile and it was taking two weeks longer than expected. And so we called up the port of Tacoma and said, "Hey, I got these seven containers coming in. We think they're gonna take two weeks to get through the port of Tacoma." Is that true? Why is it taking so long to get through the port of Tacoma? And the guy was, felt like it was in trouble, because how did you know we were sure to chassis? It was the overweight, overweight chassis in particular. We're handling heavy shipments. There was a queue of two weeks waiting to get through and he was, "Hey, wanna know how? Did you guys know?" We had kind of kept that secret, but the platform didn't really know that. It was able to look at all of the attributes, identify it as the, what was deterministic about that slow move. And I think I cheated a little bit and looked at some of your, you had done a study about a bimolar transit. - Oh my gosh. - Out of Rotterdam. - Yeah, yeah. - And we would shred that with, that's, you need to know what the composition is. That's two, that's probably a fortnightly service where there is a barge in one week and a truck in the second week. And that's what's causing that. So we would want more, we would suck in more attributes to find the determinants of those attributes to know why. - You know what John's referring to was a study we did, gosh, 15 years ago, where we had looking at, it was for a large, auto manufacturer, not mistaken, although we probably didn't disclose that. And when you look at the transit time coming into, I want to say port of Baltimore, but I could be wrong. It was bimodal. So we had, some things coming in like 14 days and some in like 21 or 28, big bumps. And what we found out, if I recall correctly, is that they were two different strings. One would come from the south up and one would come from the north down. And so they didn't know, and in their data, they just looked at the average, which never occurred. right? So they went that way. So yeah, you're right. Die. and that's one of the big things I think that AI really helped me. - Yeah, and if you connect exactly that exact model to large language models, instead of the large trade models, you would have captured string to be deterministic, direction of the string to be deterministic, and as you get all those additional attributes, now you load that up and you say, "Okay, I'm leaving, but now you know to ask or add or to it on what string." - Right, right. So let me ask you just a basic question for software, as it's developed, 'cause I'm used to software being a decision support tool to help the human. We're talking about agentic where the human's kind of removed from it, but it seems like there's two models where a system can push a solution and say, "Here's what you should be doing. You should route to Vancouver instead of going to Port of LA or it's a poll system where you let the user go in and pull the answer out." Which do you think Lognet is, or is it both? Because it brings up part of what's our autonomous model. So we look at the autonomous model as a layer that we added probably over the last four years called movement planning. And it is broken down into strategic tactical, operational, and physical planning. And those are time windows. Strategic is a whole year. Tactical is 10 weeks. Operational is 2 weeks. Physical is, you're actually moving the physical goods. And each plan, the user interacts by setting up the plans at the highest level. So that includes, you know, where am I going to source my goods from? Where am I going to ship? What's my rates going to look like? What kind of modes am I going to use? First mile, last mile? That gets all embedded into the annual plan. And then the system builds virtual truck manifests for all those different modes. Or airway bills. And then it plans that out for the whole year at excruciating detail using the same. If you use that model of the large trade model, now you've got all those attributes that you're being very specific about. And you flow that out for the whole year. So when the 10, 10, the tactical model comes along, it's going to compare with what it knows now, how many orders did you actually issue? How much are you actually sourcing? And then that 10 week forecast needs to go out to your service providers. Because now you say with a better confidence, you know, here's what I think I'm going to be doing in 10 weeks. Are you going to have that capacity? And that flows into the two week operational, which is I'm going to make my reservation for equipment or capacity. And that flows them flows into the physical. And we're creating delivery orders. We're actually doing all the work of moving the goods. So where is AI or automation having the biggest impact? My guess would be as we get more real time, you want to be more automated for the planning. You might want to have more human judgment in there. Is that, do you agree or is that different? I think it's more, it's like, you know, driving a Tesla, the more that it knows it can do. So you could, if you have a detail on a annual plan, you're just getting, it's guard rails. You can eliminate 80, 90% of the hands on that you would normally do. The great thing that the AI does is it very easily takes the happy path. And you can automate that tremendously. And I think as companies take that happy path and then deal with the outliers and the desired outcomes they would like to have with the outliers as they think that's what your manager of the future is going to be doing. Taking a look at the happy path, AI is going to slay it. But let me help me out here. What's a happy path that sounds so is that like the, when everything works path? Yes. Yes. So that's the use case that says it's a standard use case. So you get your top 80% defined in use cases in advance that you can flow the whole way. And then you've got these outliers like, uh-oh, my primary and secondary carrier fell through. How do I do a spot move? What if it's spot air spot ocean, a spot coming out of a restricted area. That's a restricted commodity and you get into these outlier use cases that you have to think about maybe what they what outcome you'd want to. Yeah. Yeah. Sometimes I run into I'm curious if you run into this where when you talk to people about a solution, they naturally go to corner cases to the extreme. Well, what if this happens rather than trying to identify that what you're calling the happy path? Because usually when I talk to them, they say, well, what about this? They focus on the corner cases instead of looking at the what 70 80% of the moves will go in streamline that first. Do you have that same situation? Absolutely, but we really push the happy path because is that trait. It all is happy path. A log net trait. I hear a lot. It's used a lot, but we have had. I hope we have happy customers. We work super hard to make them happy. Failing that to look at that happy fast and excruciated detail is why the project is going to you want to slay the happy path. You want 90% of your process working well. And you know, it's good. If it's going to be hands off, it's got to be hands off. There's so much inertia around going to autonomous solutions. People do not want to give up their spreadsheets. People do not want to. So the organizationally, there's a lot of change management that has to go into saying, hey, we're going to do this well. And the first thing we're going to focus on is the happy path. And that's got to give you a situation. We see a lot of DAT because we're mainly domestic trucking. That's where my world. I know you're much more global and end. But the happy path for truck load transportation is you do an RFP, you populate a routing guide with the in a TMS load comes in match to the TMS goes contract and goes. And that's a happy path. And it works 80% of the time. Sometimes 95% of the time during a real soft market like it is now. But one of the problems that we found is the ones that failed that that didn't go through the routing guide shouldn't have gone through it anyway. So they should have been identified way in advance. They don't even bother going through the routing guide because it won't work. And so something what you do strategically is limiting or placing the limits on what you can do operationally. Do you see the same thing that sometimes planning too much you might want to have something that you don't want to plan some of those you want to leave them for dynamic planning at down at the operational side. Does that make sense? Yeah, absolutely. So we have a mode checks as soon as a purchase orders issued. Does it make sense to can I get from point A to point B some of the basics of you have an origin, a factory origin and you have everything known about that origin. And you have supply destination points where you have either a consumer and you have a network to connect those two together all your origins and destinations. If you don't have those three pieces. Stop right there. If you've at a schedule. So the next layer is do you have a schedule that's going to meet your commitment levels for that. If you don't pass that threshold stop do not pass go if you have alternates but now your prices increasing set your price limit and hit the eject button and throw an exception. That makes sense. That makes sense. So I've used a lot of your time. I got one last question. And so we're at 2025. So up until the pandemic everyone would say future plans were in the year 2020 right everything was 2020. And now we're halfway to 2030. So what do you see in 2030 what what's next for lognet and the future of supply chain automation. What do you think will look different five years from now. I think that the beneficial owners of goods the importers the retailers, right in factors will increasingly take control their supply chain and enter and as much work as they can put on autonomous solutions they will. I think the tariffs are just one part of the puzzle but as people are going out and asking their factories you know what if I lift up my FOB curtain what are you paying and the margins taken by logistics providers factories trucking companies they're huge and they're going to say hey why am I paying a truck driver in India you know $200 a month and how come it ends up being that's not what I pay and if they from a social responsibility perspective from a ESG perspective they can reach out and pay just like Venmo for that truck driver in northern India I think they're going to jump at that evolution. That's interesting it's funny because as you make something more efficient you know one man's waste is another man's margin right and so as you make more efficient and things get cut something gets squeezed and the transparency I think will be fantastic you're going to have to go and pay the local tax authority so that government entities happy but the ability for you know a lot of developing countries are cash economies and a lot of that has a lot of people dip in into that bag of cash before it gets the person who uses it giving it straight to that party what a great model to be in so that the benefits of squeezing out you know maybe not the best use of that cash I think is a tremendously positive thing to get into and one that you know global companies once again into and they're going to be excited to see them raise the standard of living of all the people in the supply chain not have to wonder where it's that runoff is. Yeah I say I would say transparency now is late years ahead of what it was 10 15 20 years ago and is only going to increase I have phenomenally well John thank you I learned a lot I really appreciate talking to you today and learned a bunch of new things thanks no likewise it was a great chance to finally meet you. - Why didn't you look up an old paper? - Absolutely. You guys, tutorial optimization. But what is it? You have Samuelson at CMU. He started combing it, you know, Alan Holland. - That went to Jagger. - Yeah, yeah. - Oh yeah. - No, so same, yeah. So yeah, that was my dissertation. We have a story for another day about what happened. - I'm not gonna talk to you about that. - But anyway, John, thank you so much and everyone, stay tuned to the Truckload Market Update. - Take care. (upbeat music) (upbeat music) - This is the Truckload Market Update for 30 October, 2025. In Drive-Anne, we saw the change in Active Contract rates drops slightly 0.4%. Spot rates increased 3.9%, and the current level of replacement rates for Drive-Anne is negative 1.5%. Meaning, if a new Contract Rate comes in on average across all the data, it's about 1.5% lower than the rate that it's replacing. The Market Gap between Spot and Contract for Drive-Anne is negative 21 cents a mile, meaning Spot is still below Contract. For temp control, we saw Contract Rates drop 0.8%, Spot Rates rose 2.3%, and the current level of replacement rates is weirdly zero. So pretty much new rates, old rates are washed on average. And the Market Gap between Spot and Contract for temp control is negative 12 cents a mile. Intermodal saw Contract Rates rise 1.5%, the few Spot Rates in Intermodal increased 4%, current level of replacement rates for Intermodal is positive 2.5%, and we saw the Market Gap between Spot and Contract for Intermodal negative 3 cents a mile. Finally, for Flatbed, we saw Contract Rates drop 1.5%, Spot Rates increased 0.7%, and the current level of replacement rates for Flatbed is negative 5.7%, and the Market Gap between Spot and Contract for Flatbed is negative 27 cents a mile. Okay, what's going on? Well, Contract Rates dropped slightly for pretty much all modes, except for Intermodal, which increased by about 1.5%, but we've seen that it's been bouncing around for both Drive-Anne and temp control right around zero. It's not moving dramatically period to period. Spot Rates were up across the board, and that's a little uncommon. We've seen it, I guess, last period, it was the same, but generally it's been bouncing up and down. It's been so volatile with all the changes in tariffs and the macroeconomic conditions. And this has really, really been pretty volatile, and I expect it to continue to do so. Replacement rates were mixed with Flatbed's thing, consistently negative. It's negative 5.7%, it's been above 5% negative for about four months now. So that's been pretty consistent, temp control, Drive-Anne, are really approaching zero. They're bouncing up and down, but essentially what this means is, running a bid for Drive-Anne or temp control, you're not gonna squeeze much out of the market. The rock's pretty much been squeezed as much as a kin, and you're seeing more shippers try to protect capacity for when the market does, to rather than trying to harvest more savings. It's been three years of negative replacement rates, and I think that's finally coming to a slow, slow end. Finally, the gap between Spot and Contract, double-digit negative for almost all the modes, except for Intermodal. Intermodal is just barely negative, 'cause that one bounces up and down. There's just so little spot Intermodal saw. It doesn't really make as much sense. But overall, the market is kind of bouncing along in that trough, slowly getting up to parity, but not there yet. I don't see anything happening in the rest of this year, that would cause it to change dramatically. And that's the Truckload Market Update for 30 October, 2025. Happy Halloween. And that's a wrap for this episode of The Freightvine. The Freightvine podcast is hosted by myself, Chris Campos, and is produced and edited by DAT Freightnanalytics. For more information or to catch up on previous episodes, swing by our website at www.dat.com/resources/freightvine. And don't forget to hit that subscribe button wherever you listen to podcasts. And hey, why not drop us a review while you're at it? If you have any feedback or questions about what you've heard or suggestions for what you'd like to hear in the future, please send an email to me at [email protected]. That's [email protected]. And finally, a big thanks from all of us at DAT for tuning in. We hope you learned something new and you come back again. (upbeat music) (upbeat music)